Energy Landscape (To be updated)

The behaviour of many systems of interest in chemical physics (including soft matter and biomolecules) is characterized by the presence of competing interactions.  Consequently, the topography and topology defined by the potential energy function are complex, with an exponential number of local minima separated by energy barriers.  Energy landscape methods compute the kinetic, thermodynamic and structural properties of such systems based on a knowledge of the stationary points of the potential energy function.  This approach has the ability to circumvent some persistent limitations of conventional simulations, making it a powerful tool with a wide range of applications.

Together with the group of D.J. Wales (Cambridge) and S. Somani (Janssen R&D) we develop novel energy landscape methods for efficient structure predictions and binding free energy calculations of biomolecules. The basic idea is very simple. Atoms which are not central to the biochemistry of the problems are rigidified. We are currently analysing the effects of such rigidification on the potential energy surface. Preliminary results show that this is a promising approach. We tested our method on small peptides (figure; left), as well as for human aldose reductase (figure; middle). We typically gain a computational speed up by an order of magnitude (or more).

We were part of the "Simulations of Self-Assembly" programme grant and still continue to contribute to the project. Our interests are to study the self-assembly of colloidal clusters / rigid body building blocks (figure; right). Many of the projects are also in collaboration with D. Chakrabarti in Birmingham.

Our group contributes to the development of the GMIN, OPTIM, and PATHSAMPLE software.